Spatial Econometric Analysis Using GAUSS 10 Kuan-Pin Lin Portland State University
Spatial Panel Data Models The General Model
Spatial Panel Data Models Assumptions Fixed Effects Random Effects Spatial Error Model: A=I or =0 Spatial Lag Model: B=I or =0 Panel Data Model: A=B=I
Spatial Panel Data Models Example: U. S. Productivity (48 States, 17 Years) Panel Data Model ln(GSP) = + ln(Public) + 2 ln(Private) + 3 ln(Labor) + 4 (Unemp) + u + v Spatial Lag Model ln(GSP) = + ln(Public) + 2 ln(Private) + 3 ln(Labor)+ 4 (Unemp) + λW ln(GSP) + u + v Spatial Error Model ln(GSP) = + ln(Public) + 2 ln(Private) + 3 ln(Labor) + 4 (Unemp) + W e e u + v Spatial Mixed Model ln(GSP) = + ln(Public) + 2 ln(Private) + 3 ln(Labor) + 4 (Unemp) + λW ln(GSP) + W e e u + v
Model Estimation Based on panel data models (pooled, fixed effects, random effects), we consider: Spatial Error Model Spatial Lag Model Spatial Mixed Model Model Estimation Generalized Least Squares (IV/GLS) Generalized Method of Moments (GMM/GLS) Maximum Likelihood Estimation
Spatial Lag Model Estimation The Model: SPLAG(1) OLS is biased and inconsistent.
Spatial Lag Model Estimation Fixed Effects
Spatial Lag Model Estimation Fixed Effects: IV or 2SLS Instrumental Variables Two-Stage Least Squares
Spatial Lag Model Estimation Random Effects
Spatial Lag Model Estimation Random Effects: IV/GLS Instrumental Variables Two-Stage Generalized Least Squares
Spatial Lag Model Estimation Random Effects: IV/GLS Feasible Generalized Least Squares Estimate v 2 and u 2 from the fixed effects model: FGLS for random effects model:
Spatial Error Model Estimation The Model: SPAR(1) Fixed Effects Random Effects
Spatial Error Model Estimation Fixed Effects Moment Functions
Spatial Error Model Estimation Fixed Effects The Model: SPAR(1) Estimate and iteratively: GMM/GLS OLS GMM GLS
Spatial Error Model Estimation Random Effects Moment Functions (Kapoor, Kelejian and Prucha, 2006)
Spatial Error Model Estimation Random Effects The Model: SPAR(1) Estimate and iteratively: GMM/GLS OLS GMM GLS
Spatial Mixed Model Estimation The Model: SARAR(1,1)
Spatial Mixed Model Estimation Two-Stage Estimation Sample moment functions are the same as in the spatial error AR(1) model. The efficient GMM estimator follows exactly the same as the spatial error AR(1) model. The transformed model which removes spatial error AR(1) correlation is estimated the same way as the spatial lag model using IV and GLS.
Spatial Mixed Model Estimation Fixed Effects The Model: SPARAR(1,1)
Spatial Mixed Model Estimation Fixed Effects Estimate and iteratively: GMM/GLS IV/2SLS GMM GLS
Spatial Mixed Model Estimation Random Effects The Model: SPARAR(1,1)
Spatial Mixed Model Estimation Random Effects Estimate and iteratively: GMM/GLS IV/2SLS GMM GLS
Example: U. S. Productivity Baltagi (2008) [munnell.5]munnell.5 Spatial Panel Data Model: GMM/GLS (Spatial Error) ln(GSP) = + ln(Public) + 2 ln(Private) + 3 ln(Labor) + 4 (Unemp) + =ρW + e, e = i u + v Fixed Effectss.e Random Effectss.e 0.202* *0.021 3 * *0.025 4 * *0.001 0 *0.136 ρ0.578* *0.060
Example: U. S. Productivity Baltagi (2008) [munnell.5]munnell.5 Spatial Panel Data Model: GMM/GLS (Spatial Mixed) ln(GSP) = + ln(Public) + 2 ln(Private) + 3 ln(Labor) + 4 (Unemp) + λW ln(GSP) + =ρW + e, e = i u + v Fixed Effectss.e Random Effectss.e 0.185* *0.022 3 * *0.026 4 * *0.001 0 *0.174 λ0.093* *0.015 ρ0.488* *0.059
Another Example China Provincial Productivity [china.9]china.9 Spatial Panel Data Model: GMM/GLS (Spatial Error) ln(Q) = + ln(L) + ln(K) + =ρW + e, e = i u + v Fixed Effectss.e Random Effectss.e ρ
Another Example China Provincial Productivity [china.9]china.9 Spatial Panel Data Model: GMM/GLS (Spatial Mixed) ln(Q) = + ln(L) + ln(K) + W ln(Q) + =ρW + e, e = i u + v Fixed Effectss.e Random Effectss.e λ ρ
Maximum Likelihood Estimation Error Components Assumptions Fixed Effects: Random Effects:
Maximum Likelihood Estimation Fixed Effects Log-Likelihood Function
Maximum Likelihood Estimation Fixed Effects Log-Likelihood Function (Lee and Yu, 2010) Where z* is the transformation of z using the orthogonal eigenvector matrix of Q.
Maximum Likelihood Estimation Random Effects Log-Likelihood Function
Example: U. S. Productivity Baltagi (2008) [munnell.4]munnell.4 Spatial Panel Data Model: QML (Spatial Lag) ln(GSP) = + ln(Public) + 2 ln(Private) + 3 ln(Labor) + 4 (Unemp) + λW ln(GSP) + , = i u + v Fixed Effectss.e Random Effectss.e 0.187* *0.025 3 * *0.029 4 * * 0 *0.166 λ0.275* *0.029
Example: U. S. Productivity Baltagi (2008) [munnell.4]munnell.4 Spatial Panel Data Model: QML (Spatial Error) ln(GSP) = + ln(Public) + 2 ln(Private) + 3 ln(Labor) + 4 (Unemp) + =ρW + e, e = i u + v Fixed Effectss.e Random Effectss.e 0.205* *0.023 3 * *0.027 4 * *0.001 0 ρ0.557* *0.033
Example: U. S. Productivity Baltagi (2008) [munnell.4]munnell.4 Spatial Panel Data Model: QML (Spatial Mixed) ln(GSP) = + ln(Public) + 2 ln(Private) + 3 ln(Labor) + 4 (Unemp) + λW ln(GSP) + =ρW + e, e = i u + v Fixed Effectss.e Random Effectss.e 0.191* *0.023 3 * *0.027 4 * *0.001 0 *0.212 λ ρ0.455* *0.038
Another Example China Provincial Productivity [china.8]china.8 Spatial Panel Data Model: QML (Spatial Lag) ln(Q) = + ln(L) + ln(K) + W ln(Q) + = i u + v Fixed Effectss.e Random Effectss.e λ
Another Example China Provincial Productivity [china.8]china.8 Spatial Panel Data Model: QML (Spatial Error) ln(Q) = + ln(L) + ln(K) + =ρW + e, e = i u + v Fixed Effectss.e Random Effectss.e ρ
Another Example China Provincial Productivity [china.8]china.8 Spatial Panel Data Model: QML (Spatial Mixed) ln(Q) = + ln(L) + ln(K) + W ln(Q) + =ρW + e, e = i u + v Fixed Effectss.e Random Effectss.e λ ρ
References Elhorst, J. P. (2003). Specification and estimation of spatial panel data models, International Regional Science Review 26, Kapoor M., Kelejian, H. and I. R. Prucha, “Panel Data Models with Spatially Correlated Error Components,” Journal of Econometrics, 140, 2006: Lee, L. F., and J. Yu, “Estimation of Spatial Autoregressive Panel Data Models with Fixed Effects,” Journal of Econometrics 154, 2010: